Wednesday, 29 November 2017

BT Heads to Bristol As It Weighs 5G Plans, Costs and competition strategy

Eager to figure out how a 5G network behaves in the "real world," what applications actually need a next-generation architecture, and how a 5G network might be managed, BT is teaming up with Nokia and the "Bristol Is Open" initiative for a two-day live trial in March 2018 that will involve up to 5,000 members of the public.

Bristol, a city in west England that is home to more than 450,000 people and countless innovative technology and media companies (such as Blu Wireless Technology, Venture Photonics, Zeetta Networks), has for some time been a live test bed for various transport network, wireless and IoT applications under the auspices of the Bristol Is Open initiative, with BT being one of countless R&D partners.

Now the UK operator is seeking to gain significant insights into the 5G world from the two-day experiment that will take place in Bristol city center. Pre-standards 5G, 4G, 3G and WiFi access networks will all be operational as multiple applications and services are tested over various spectrum bands (3.5GHz, 26-28GHz and 60-70GHz) using a variety of end user devices, including two pre-5G and still sizeable prototype devices (that will gain their mobility by being carted around on golf buggies, apparently). Li-Fi capabilities will also be tested in some indoor environments. (See Eurobites: WiFi's Lightbulb Moment.)
With just two 5G prototype devices, the number of fully end-to-end 5G service tests will be limited, though versions of some potential 5G applications will be demonstrated and trialled using LTE-Advanced Pro (4.5G) radio access, while 5G NR connectivity will also be tested as backhaul connections.

The trial will make use of Bristol Is Open’s existing optical transport and data gathering architecture, spectrum from BT (which of course now includes mobile operator EE) and network infrastructure elements -- massive MIMO antennas, pre-standard 5G NR small cells, edge computing platforms, core network systems and more -- from Nokia.

Talking at a media briefing Monday morning, BT's Chief Network Architect Neil McRae noted that BT needs to be able to "do more for less, with less" in the future, and that it needs practical experience of how next-generation technologies will work and interact when loaded with real user traffic. "We want to look at this from an operational perspective... how to manage network slicing... what tools we will need to manage 5G networks that we don't already have and looking at how 5G [capabilities] might be used in the real world beyond enhanced mobile broadband connectivity."

Identifying and implementing the appropriate analytics capabilities will be critical to BT's 5G plans, noted McRae, as well as figuring out "how to get the information out of the network. How to improve telemetry is a key area of focus for BT and then figuring out what information needs to be processed and analyzed in real time and what can be stored and analyzed later. Information is the starting point for 5G -- visibility gives us control."

BT's Chief Architect Neil McRae (seen here talking to Light Reading's Liz Coyne earlier this year) believes that 'Information is the starting point for 5G -- visibility gives us control.'
BT's Chief Architect Neil McRae (seen here talking to Light Reading's Liz Coyne earlier this year) believes that "Information is the starting point for 5G -- visibility gives us control."

Neil McRae also stressed that it's important to look at the technology as the enabler and be driven primarily by the potential use cases in a 5G world, "and the most powerful use cases come from collaborating with industry... we need to work with government, industry, enterprises to provide everyone with a better understanding of what impact 5G will have on businesses, their partners and their customers." Automation and robotics are at the heart of the applications to be tested, and will involve interactions between humans and robots, between robots and robots, with connected cars and also encompass "immersive experiences," including some involving real-time gaming.

That's one of the incentives to be involved in the Bristol experiment -- collaboration and feedback. "We are looking at Bristol as a way to engage in applied research," noted Paul Crane, head of mobile, wireless and network services R&D at BT's Technology, Service & Operations (TSO) division. Having worked on 5G capabilities in the labs, BT wants to know "what we need to put in place" to enable a slew of 5G applications that require high capacity, high availability and low latency.

Crane noted that his team has three particular areas of 5G R&D in terms of applications: Low-latency services, such as remote medical diagnostics; Virtual and augmented reality, particularly where VR services pertain to sporting events and live sporting broadcasts (BT, of course, now runs multiple sports TV channels in the UK); and drone control, focused on how automated drone missions can help the emergency services and make remote site visits more efficient and effective.

He is also looking at: How artificial intelligence and machine learning tools, which BT already uses in its network maintenance processes, could be used in automated 5G management processes; the integration of fixed and mobile network infrastructures all the way to the access edge; and how functions can be more distributed.

BT has already been examining distributed and edge computing capabilities and has already deployed distributed content caches around the UK. McRae noted that BT is now seeing more value in its local exchange (central office) facilities than it did several years ago, as the ability to add compute functionality in those locations "gives us a huge asset to build on," he noted.

But that doesn't mean that all of BT's circa 5,500 local exchanges will become distributed data center facilities: McRae believes about 2,000 will be useful in the long run to house edge computing as well as broadband access systems and "putting edge compute resources in about 100 locations would probably be enough to start."

And while BT has found the CORD (Central Office Re-architected as a Datacenter) developments instructive, "we have a different set of needs" that are being met with developments within Telecom Infra Project (TIP) and through work BT is undertaking with Dell. (See BT, Dell EMC Team on Disaggregated Switching R&D.)

So what happens after the brief Bristol trial? A great deal of information will be gathered from the two days, noted Bristol is Open CTO Dimitra Simeonidou, an optical technology specialist and professor of high performance networks at Bristol University. She said that services would be run over 4G and pre-5G connections and data gathered to note the difference in performance, network impact and so on. Information will be shared from the experiment with the broader community she added: "We are looking to create a 5G asset for the whole of the UK," she said, though it seems very likely that critical, detailed performance insights might not go much further than the key participants.

For BT, McRae will be looking for insights that can feed into the operator's 5G network planning and cost models. As expected he wasn't willing to provide a figure for how much capex BT would need to build out a nationwide 5G network with distributed NR radio access, edge computing, real-time analytics and ultra high-speed backhaul, but there is, he notes, "a five-year plan, maybe not as good as it might be currently, but we're getting closer to a more defined plan. There's still a lot to understand about how 5G will feed into BT's overall service and architecture plans and we're still looking at that."

Those plans don't include new spectrum license costs, around which there is still a great deal of uncertainty -- not just around the price but also the timing, as BT is currently involved in a dispute related to potential spectrum allocation rules. (See Eurobites: UK Regulator Accuses BT & 3 of Erecting 5G Roadblocks and Hot Air: UK 5G Spectrum Dispute .)

Neil McRae dismissed the idea that the spectrum dispute might impact BT's 5G plans. "Anyone who thinks that... doesn't know what they're talking about," he proclaimed. If you think that sounds combative, try asking him if he's looking forward to 6G networks!

Sunday, 26 November 2017

Google developing customer service teams?

The obvious approach to Customer Service that is more up Google's alley, is to simply develop a customer service chat interface, which it can easily integrate with Google Classroom or Google Assistant and sell it to educational establishments and other  businesses to use. Eventually, after enough data is going to be massed, Google could likely handle all customer interactions on its own (maybe with some live people monitoring the conversations but only stepping in under rare circumstances).

This would save Google the massive expense of developing a huge customer service department, training programs, and potentially outsourcing the chat interface software.

Some bulleted lists detailing the pros and cons of each strategy are found below:

Starting A Live Customer Service Department: Allow for product support on their own products.

Create another direct customer interaction point besides stores (and one that is potentially much more impactful for a digitally-based company).

Allow collection of data on customer service interaction expediting the proliferation of an AI bot version.

Create A Customer Service Chat Interface: Allow them to create a new interface for their own use, that is also a marketable B2B product for the G Suite they already market to schools and businesses and Google Classroom for the rest of the world.

Depending on terms of use, it could collect data from the clients allowing an even more robust set of data for creating the AI version.

What About Google doing Both?

By doing both of these strategies in conjunction, there are many synergies: One of the most obvious is that it would increase justification to create an in-house version of the live chat interface (whether from scratch or a ported version of Hangouts, Duo, or any number of other chat interfaces they have already created like the little used Google+).

This would reduce the huge cost from buying the SaaS from a company like Livechat. The fact remains that it would then be a Google product to fold into G Suite or sell on its own with appeal to enterprises would be a great up-sale for cloud services or other products.

Additionally, the costs could also be offset for employees they hired to do the customer service via the interface while honing the AI by potential revenues from selling the software.

The main reason to do this would be that this is one area they lack behind many companies they compete with like Apple, Microsoft and Amazon. Further, many of Google's apps are basically unsupported. For example, if a business or school migrates from Google G Suite to a competitor, there is nobody at Google to help people transfer their documents to another Google account, let alone the competitor. There is nobody to talk to when you are having trouble with the software. The most you and I can do is file a bug report requesting a change in features. NOT good enough in our Healthy  Mindful MenteSanas future plans. We see such a clear view from Project Loon's 12 mile high magination. It's "just" a matter of bringing the 21st century, New Millennium customer service idea for Alphabet down to earth.

Tuesday, 21 November 2017

Apple leans heavily upon Intel's 5G CHIPSET

Apple is leaning heavily toward Intel’s flavour of 5G for a future iPhone. The iPhone maker’s engineers have been engaged with Intel counterparts for early work on 5G, the upcoming technology for next-generation wireless broadband, while dialogue between Apple and the dominant modem supplier in the industry, Qualcomm, has been limited.

Qualcomm’s 5G modem chips offer more specialized carrier features like “Uplink Carrier Aggregation“ but many will not be widely adopted by carriers. Apple engineers believe that Intel’s 5G modem will fit its requirements for a future iPhone. Intel has lagged far behind Qualcomm in the modem market, but has a small army working on 5G, numbering in the multiple thousands. The initiative to provide the 5G modem for the iPhone is now considered a “must-win” for Intel. Intel first announced its 5G Modem at CES 2017, and said today in a press release that it has “successfully completed a full end-to-end 5G call based on its early 5G silicon … a key milestone in its development.” The completion of the modem aligns with Apple’s plans for a 5G iPhone to arrive in 2019 or 2020. A 5G iPhone will be capable of connection speeds of a gigabit per second and beyond, which could radically change the way we compute, communicate, and consume content using the device. The move to 5G is technically complex, requiring lots of coordination between the phone maker and potential vendors far in advance of official vendor selection.


Qualcomm has been supplying modems for the iPhone since 2011, and has a far more mature and full-featured 5G modem. After a big R&D investment, the company announced the world’s first 5G modem, the Snapdragon X50, back in October 2016. Intel, after its own gargantuan effort, won a small share of the Apple modem business starting with 2016’s iPhone 7. The company has been willing to customize its products for Apple in ways that Qualcomm likely would not. Apple has used the Intel modems in its iPhones for the T-Mobile and AT&T GSM/LTE networks, and, because the Qualcomm modems are especially well suited to CDMA networks, has put them in iPhones for Verizon and Sprint.

That dynamic may now change as the carriers transition their networks to 5G. Verizon, for one, has said it plans to no longer require that new devices connect to older CDMA networks starting in late 2018 or early 2019. When Verizon does that, Sprint, US Cellular, and a few other overseas operators will still support CDMA, but the CDMA ecosystem will begin to collapse.

Without a CDMA requirement, Apple has one less reason to stick with Qualcomm, which invented the widely used version of the technology.

A Strained Relationship Is EVIDENT today.

Meanwhile, Apple’s relationship with Qualcomm has soured over the past year amid an increasingly serious legal dispute over the patent licensing fees paid to Qualcomm by Apple suppliers. Also, Qualcomm is now the subject of a monster acquisition attempt by rival chipmaker, Broadcom, but has already rejected an eye popping $103 billon bid, which it said undervalues its business.

All of this seems to point toward the possibility, (or is it probability?) of Apple looking to Intel as the sole provider of 5G modems for the future iPhone. The end game is obviously to build the Intel modem onto an integrated system-on-a-chip (SoC) that would also contain the CPU, GPU, and most other iPhone components. The SoC would be co-designed by Intel and Apple plus most importantly the chip would be fabricated at an Intel facility. Some believe Apple’s full embrace of Intel could happen even sooner. The Wall Street Journal recently cited unnamed sources saying that Apple is already building iPhone prototypes that use Intel 3G/4G modems only. The report named a smaller chipmaker, Taiwan-based MediaTek, as a possible (if somewhat unlikely) second supplier.

Qualcomm would not go on record for this story. (Suppliers are required to sign strict non-disclosure agreements that prohibit talking publicly about their business with Apple.) Apple didn’t respond to a request for comment. Intel did however provide a telling statement: “While we do not comment on customer products, as evidenced by our news today Intel is making great momentum on our 5G roadmap to accelerate the adoption of 5G.”

Thursday, 2 November 2017

More on KRACKS, the WiFi WPA2 attack technique reads info that was previously assumed to be safely encrypted.

The recent discovery of serious weaknesses in WPA2, a protocol that secures all modern protected Wi-Fi networks, has been in the headlines. An attacker within range of a victim can exploit these weaknesses using key reinstallation attacks otherwise known as KRACKs. 

Concretely, any attacker can use this novel attack technique to read information that was previously assumed to be safely encrypted. This can be abused to steal sensitive information such as credit card numbers, passwords, chat messages, emails, photos, and so on. 

The attack works against all modern protected Wi-Fi networks

Depending on the network configuration, it is also possible to inject and manipulate data. For example, an attacker might be able to inject ransomware or other malware into websites.

The weaknesses are to be found in the IEEE 802.11 Wi-Fi standard itself, and not in individual products or implementations. Therefore, any correct implementation of WPA2 is likely affected. To prevent the attack, users must urgently update their affected products as soon as security updates become available. If ANY of your devices support Wi-Fi, it is most likely affected

During initial research, it was discovered that Android, Linux, Apple, Windows, OpenBSD, MediaTek, Linksys, and many others, are all affected by some variant of the attacks. 

For more information about specific products, consult the database of CERT/CC, or contact your vendor.

The research behind the attack will soon be presented at the upcoming Computer and Communications Security (CCS) conference, and at the Black Hat Europe conference. 

A very detailed research paper can already be downloaded. The recommendation from MenteSanas is not to delay acting on this risk any further. 

Wednesday, 1 November 2017

Alphabet's (Google's) routing blunder sent Japan's Internet dark last Friday!

An indeterminate but supposedly small number of Google Docs including MenteSanas users on Tuesday found that their essays, reports, school assignments, tracts, and manifestos had run afoul of Google's terms of service and had been made inaccessible.

Some users reported being unable to share their documents; others said their documents could not be viewed in Google Drive; and a few of us claimed their work had been lost, though we're told what was lost has fortunately been found again.

Several hours ago, Bhaskar Sunkara, founder of Jacobin magazine, via Twitter said an article on Eastern Europe's post-socialist policies had vanished from his Google Drive space due to a terms of service (ToS) violation.

Rachel Bale, a reporter for National Geographic, said a draft of a story about wildlife crime had been frozen for a ToS violation.

And Jason Heppler, an assistant professor of history at the University of Nebraska at Omaha, posted a screenshot showing that a requested file had been deleted from Google Drive.

Similar tales litter the Google Docs Help Forum.

The incident prompted reiterations of longstanding concerns about the downside of cloud-based services, namely that files stored remotely can be swept away at any time for any reason. And it comes at a time when Google and its peers are under scrutiny in the US for not knowing more about those who share content and pay for ads on social platforms.

Google routing blunder sent Japan's Internet dark on Friday

Google offers Docs and Drive under the usual rules, which disallow abusive or illegal content. In most circumstances, it does not scrutinize private content stored on its servers, though it does have automated systems in place for detecting illegal images, at least in Gmail.

YouTube also has a system for detecting copyrighted content.

For shared content like Docs files, much of the ToS policing has been foisted upon other users, who are invited to flag material they deem inappropriate.

It's that flagging mechanism that went haywire, we're told. Shortly after noon Pacific Time, Google acknowledged its errant flagging frenzy and attributed it to bad code.

"This morning, we made a code push that incorrectly flagged a small percentage of Google Docs as abusive, which caused those documents to be automatically blocked," a Google spokesperson said in an emailed statement to The Register. "A fix is in place and all users should have full access to their docs."

Google's spokesperson said that protecting users from malware and abusive content is part of the company's strategy for keeping users safe online and apologized for the disruption while promising to implement safeguards to prevent its safeguards from getting out of hand.

MenteSanas are no longer in Stealth Mode

From today, November 1st, 2017 we at MenteSanas are no longer in Stealth Mode as soon as the "Hangouts Meet hardware kits" arrive, as we will have been patiently waiting several months. 

They go on sale in the U.S.A. and Canada, the U.K., Ireland, Finland, France, Norway, Sweden, Spain, Australia, Japan, and even New Zealand.